The long-term goal of this research is to support the ability of Veterans with dementia to remain living in their homes and communities (age in place) by ensuring continuous access to the highest possible quality of care, regardless of where care is sought. A key element of high quality care in Veterans with dementia is safe and effective medication prescribing. Given their complex healthcare needs, Veterans with dementia require considerable care coordination across providers and systems of care to avoid inappropriate or over-prescribing of medications. With implementation of the Medicare Modernization Act, Veterans with VA pharmacy benefits became eligible in 2006 to enroll in Medicare's prescription drug benefit program (Medicare Part D). While dual VA/Medicare Part D eligibility may offer greater flexibility and access to medications, dual users of VA/Medicare drug benefits may be vulnerable to suboptimal prescribing resulting from poor care coordination across fragmented systems of care Despite these potential risks, we know very little about factors that drive different patterns of dual VA-Medicare Part D use in Veterans with dementia, or what impact dual use has on medication safety and outcomes. We propose a mixed-method (quantitative and qualitative) study in order to address these important gaps. The specific aims are to: (1) describe patterns of outpatient medication use across drug benefit user groups (VA-Only, Medicare Part D-Only, Dual Users) and identify risk factors for dual use; (2) evaluate the effects of dual use on suboptimal medication use (potentially harmful medications in the elderly, drug-disease interactions, drug-drug interactions, medication surplus) and Veteran outcomes; and (3) explore Veteran, caregiver, and VA provider perspectives on key reasons for dual use, perceived risks, and potential care coordination solutions. The quantitative phase of this mixed-method study will involve analyses of VA and Medicare data in ~147,000 dually enrolled VA patients with dementia. We will link inpatient, outpatient, nursing home, vital status, and medication data from VA and Medicare for a national cohort of veterans with dementia who are both active VA users and enrolled in Medicare fee-for-service. Our analytic approach involves propensity score matching and instrumental variables to address potential selection bias associated with dual VA/Medicare Part D use, and structural equation modeling to determine if suboptimal prescribing mediates the relationship between dual use and Veteran outcomes. The qualitative phase will involve semi-structured interviews with dual-using Veterans with dementia patients/informal caregivers and their VA primary care provider to assist in interpreting quantitative results, and gain additional insights about reasons for dual use and potential solutions. Together, these aims will provide VA and the Centers for Medicare & Medicaid services with information critical to understanding patterns, influences, and outcomes associated with dual use in Veterans with dementia, filling a crucial gap in evidence needed for appropriate policy and clinical interventions.